probability panel
Ipsos
Axios-Ipsos American Health Index, Wave 2 · May 12–15, 2023 · US national
Methodology
- Pollster
- Ipsos commercial
- Sponsor
- Axios
- Commissioned by
- Axios
- Fielded
- May 12–15, 2023
- Population
- adults
- Sample size
- 1,095
- Sample frame
- probability panel
- Panel
- Ipsos KnowledgePanel
- Mode
- online - self-administered web survey of the probability-based KnowledgePanel; address-based sampling via the USPS Delivery Sequence File
- Margin of error
- +/- 3.1 percentage points at the 95% confidence level for the entire sample of adults (design-effect adjusted)
- Weighting
- Weighted to adjust for gender by age, race/ethnicity, education, Census region, metropolitan status, household income, and party ID
- Geography
- US national
- Base notes
- The published topline reports wave-level totals only; no party or demographic crosstabs are printed. Ipsos's own writeup states that party gaps on these items run 35-40 percentage points or more, but does not publish the underlying figures.
Sources
Verification note. Full 17-page Axios-Ipsos Wave 2 topline retrieved and text-extracted. NOTE ON A PUBLISHED DISCREPANCY: Axios's own write-up of this poll reported the parent/child-abuse item as 38% agree / 36% disagree, whereas the Ipsos topline prints 36% agree / 38% disagree. The Ipsos topline figures are used here as the primary source. This wave is the source of the widely-cited 44%/31% gender self-determination split.
Figures below are transcribed from these documents. Where they disagree with the source, the source governs.
Questions asked (3)
whether people should be able to decide their own gender identity
How much do you agree or disagree with the following statements?
How much do you agree or disagree with the following statements? People should be able to decide for themselves their gender identity
Wording notes: offers a true midpoint, taken by 24% - a quarter of adults decline to take a side on the foundational question | paired with an otherwise identical item substituting 'Teenagers' for 'People', which makes it a clean age manipulation | the response distribution is strongly bimodal: 28% strongly agree, 24% strongly disagree, with only 16% and 7% in the 'somewhat' categories
neutral — Clean, symmetric construction with a genuine midpoint - one of the better-designed identity-belief items in this corpus. Its pairing with the teenager version is the analytic payoff: support falls from 44% to 33% and opposition rises from 31% to 44% on the identical proposition when the subject is a teenager, isolating age as the decisive variable rather than the principle of self-determination.
whether teenagers should be able to decide their own gender identity
How much do you agree or disagree with the following statements?
How much do you agree or disagree with the following statements? Teenagers should be able to decide for themselves their gender identity
Wording notes: differs from the adult version only in the subject noun - a deliberate within-battery age manipulation | 'Teenagers' is unqualified as to age, covering both 13-year-olds and 19-year-olds | concerns identity self-definition, not medical intervention - a distinction the public debate frequently collapses
neutral — Neutrally worded with the same full scale as its adult counterpart. Note carefully what this item does and does not measure: it asks whether teenagers may decide their own gender identity, which is a question about self-understanding, not about access to hormones or surgery. That 44% disagree even with the self-definition version establishes that opposition to trans youth is not confined to medical questions.
whether supportive parents of transgender children are committing something like child abuse
How much do you agree or disagree with the following statements?
How much do you agree or disagree with the following statements? Parents supporting their transgender children with transitioning genders are committing something like child abuse
Wording notes: 'something like child abuse' is a hedged formulation - it invites agreement with a moral analogy rather than a factual claim, which lowers the bar for endorsement relative to asking whether it IS child abuse | agree/disagree format, so acquiescence bias runs toward the accusation | 'transitioning genders' is vague as to what is involved - social transition, blockers, hormones and surgery are all compatible with the wording | 24% took the midpoint
loaded_terms — 'Child abuse' is among the most severe accusations available in English and the item asks respondents to endorse it against parents, hedged by 'something like'. The hedge is what makes it loaded rather than merely blunt: it lets a respondent register moral disapproval without committing to the literal claim, so the 36% agreement cannot be read as 36% believing these parents are abusers. The term entered polling because Texas made it official policy in 2022, so its use is traceable to real policy language - but the construction still tilts the measurement.